The present invention relates to instruments for measuring optical characteristics—for example color, translucency, and/or gloss—of objects and, more particularly, to such instruments for use in dental applications.
The determination of shade or color of an object is a process that is frequently performed in the field of dentistry. To perform dental restorations of a damaged tooth, a dentist visually compares the color of the tooth to be replaced with an assortment of shade tabs. These shade tabs are physical tabs representing the color of commercially available restorative material, such as ceramics. The tabs include the exact specification of materials needed to produce a restorative tooth with the shade of the original tooth as determined by the dentist's visual comparison. Once the dentist finds a shade tab that matches the color of the tooth, or remaining adjacent teeth in some cases, he is in a position to manufacture the required restoration. This process, however, is very time consuming and quite subjective, and frequently results in poorly shaded restorations.
In the field of dentistry, intraoral cameras are frequently used to acquire images of teeth and determine treatment plans for cavities and other mechanical reconstruction. These cameras are designed to be versatile and able to collect measurements in tight places often found in the mouth; however, they do not preserve the color fidelity—that is, they do not collect the true color—of the object measured.
Some dentists attempt to use intraoral cameras to assist in the shade determination process. Unfortunately, conventional intraoral cameras suffer two problems: distance sensitivity due to illumination geometry and color discrimination error due to sensor limitations.
With regard to the first problem, intraoral cameras typically use fiberoptic illumination to reduce the size of the handpiece. Such a device is disclosed in U.S. Re. Pat. No. 36,434 to Hamlin et al, reissued Dec. 7, 1999. The goal of Hamlin, and most intraoral cameras, is to provide a small measuring tip on a handpiece that can be used to probe hard to reach areas in a mouth. Although fiber-optic illumination is useful for providing high levels of illumination and is compatible with small measuring probe tips, a drawback of any small illumination-source that illuminates a larger area is that the projected beam must be divergent light. The intensity of a divergent beam is governed by the inverse square law given below:
                    I        =                              D            2                                              (                              D                +                                  Δ                  ⁢                                                                          ⁢                  D                                            )                        2                                              (        1        )            where I is intensity, D is the distance from the illumination source, and ΔD is an increase in distance D from the light source. The concept of Equation 1 is illustrated in FIG. 1, where fiber optic source 115 projects illumination flux 112 to distances D and D plus ΔD. There, the intensity of flux 112 at distance D, according to Equation 1, is greater at distance D from light source 115 than at a distance D plus ΔD.
It is known that when the distance change to the illumination source is significant with respect to the distance to the source, the illumination output varies significantly, creating what is called non-uniform illumination. Particularly with objects positioned close to the fiber optics, certain regions of the object are non-uniformly illuminated because the light from the illumination source rapidly diffuses as it travels away from the source. Moreover, when multiple sources of light are used to illuminate an object, the object may be non-uniformly illuminated in different regions.
An example of non-uniform illumination of the surface of an object is understood with further reference to FIG. 1. As depicted there, a curved surface of a tooth T, slightly exaggerated for purposes of discussion, is illuminated within flux 112 projected from light source 115. Region of the tooth 113, lies distance D from light source 115, and region 114, lies distance D plus ΔD from the light source 115. As explained above, the intensity of light is greater at distance D than at D plus ΔD. Accordingly, regions 113 and 114 are not illuminated with the same intensity of light, that is, illumination is non-uniform. Sensors sensing light reflected from tooth T will collect inconsistent color information from these regions.
An example of non-uniform illumination of regions of an object with a multiple fiber optic light sources is illustrated in FIG. 2. Exemplary fiber optic light sources 120 and 122 project light fluxes 130 and 140 to illuminate the tooth T These light fluxes reflect from the tooth and are collected by an image sensor not shown for the sake of simplicity. As can be seen, tooth region 122 is illuminated primarily by light flux 140, but region 124 is illuminated by a combination of light fluxes 130 and 140. Of course, this illumination is three-dimensional even though it is depicted here in only two dimensions. Further, if more fiber optic light sources are added, the tooth is subdivided into even more regions of different illumination overlap. Given this non-uniform illumination, a color sensor, sensing the light reflected from the tooth, will invariably collect inconsistent color information from region to region. For example, what is sensed as “lighter shade” in region 122 may be sensed as “darker shade” in region 124 due to the non-uniform illumination.
With non-uniform illumination, conventional intraoral cameras critically rely on illumination source positioning which can not be maintained in practical use. This results in significant errors affecting tooth shade determination.
Other devices, specifically designed for tooth shade determination, have been proposed that use bi-directional fiber optic illumination. Such a method is described in U.S. Pat. No. 6,038,024 to Berner, issued Mar. 14, 2000. A limitation of this method of illumination is that the illuminant intensity is maximized at the intersection of the two projected beams. Often, significant portions of the measured area are not illuminated by both beams and hence have a lower and unpredictable illumination value.
Berner's non-uniform illumination is depicted in FIG. 3. A fiber optic bundle 150 is supplied with light at one end. Prior to arriving at the probe tip, the bundle is bifurcated, or divided into two bundles 152 and 154. The bundles are mechanically aimed at the target tooth T in some fixed angularity. Collimating lenses 156, 158 are often added in the path of illumination between the fiber optic bundle and the target T to lower distance sensitivity of illumination output. Each bundle generates a light flux 162 and 164 projected onto tooth T from two directions with collimating lenses 156, 158. As can be seen, fluxes 162 and 164 intersect on tooth T resulting in the intensity in region 169 being greater than the intensities in regions 167 and 171 because those regions 167 and 171, and other peripheral regions, are each illuminated by light fluxes 164 and 162 individually. The fluxes reflected from the tooth T are not shown for simplicity.
Given this non-uniform illumination, a color sensor, sensing the light reflected from the tooth, will invariably collect inconsistent color value information from region to region. For example, what is sensed as “lighter shade” in region 167 may be sensed as “darker shade,” in region 169 due to the non-uniform illumination. Moreover, with multiple light source paths, gloss artifact potential is increased. Where glare artifacts exist, the color of the target is washed out by the image of the light source itself rather than the desired tooth subject.
In addition to non-uniform illumination, today's intraoral cameras utilize color filter array (CFA) image sensors that frequently contribute to inaccurate color measurement because the filter array is applied to the image. Many intraoral cameras include color filter arrays such as red, green and blue (RGB) arrays, and cyan, magenta, yellow and green (CMYG) arrays, to name a few. Generally, these color filter arrays are made up of a multitude of adjacent elements called “pixels” (i.e., picture elements). Each pixel measures only the bandwidth of light it is designed to collect. Therefore, in a region of a image corresponding to a pixel, only the bandwidth of light specific to that pixel is displayed, even though the object measured may include other colors in that region.
The operation of and problems associated with color filter arrays are more readily understood with reference to a particular array. A few pixels of a CFA RGB sensor are illustrated in FIG. 4 as R, B, G. These R, G, and B pixels collect, capture or sense light corresponding to red, green and blue wavelengths impinging on the sensor respectively. The RGB sensor converts these collected wavelengths into electronic data and passes this data to a processor for display of a color image of the tooth on a monitor. Although RGB sensors offer a means to collect color data for a tooth, that data often is not an accurate representation of the true color or distribution of color on the tooth.
CFAs do not accurately measure color primarily because of two factors: pixel spacing separation and poor color fidelity. First, the pixel spacing separation factor may be understood with reference to the RGB sensor in FIG. 4. Each individual R, G, and B pixel in the RGB array 100 collects only one bandwidth of light reflected from a point on a tooth, for example, only red, only green, or only blue. Thus, when tooth sections 101 and 102 are illuminated and reflect light toward the RGB array and that light is detected by the corresponding G and B pixels respectively, only green bandwidths are collected by the green pixel and only blue bandwidths are collected by the B pixel. Even though section 101 actually may be blue, green, red, yellow or any other color of the spectrum, and consequently reflect the associated bandwidths, only the green bandwidth, if any, is detected by pixel G in section 101. Similarly, section 102 may be green or any other color, but those colors are not detected by the B pixel because blue is the only bandwidth that it can collect.
Accordingly, RGB sensors collect only one bandwidth for each point on the tooth even though that point may reflect many bandwidths. As a result, any measurement data for that point will include only data selectively collected by the R,G, or B pixel associated with that point. Moreover, prosthesis manufactured from this measurement data collected with an RGB sensor will not accurately reflect the true color of each point on the tooth. This phenomenon applies to all CFA sensors.
The second factor affecting color measurement is poor color fidelity of CFAs. The mass market for color sensors, in particular CFAs, is consumer electronics and video applications. The goal of such devices is to provide good image resolution, high image acquisition speed and reasonable color fidelity as needed for broadcast and personal imaging applications. CFAs are designed to be inexpensive to manufacture, to provide direct acquisition of RGB data and to provide reasonable low light performance. These design goals come at the cost of color fidelity. More specifically, today's RGB CFA collects selected wavelengths of light impinging on them, but they also incidentally collect unwanted wavelengths in the process. For example, a blue pixel of an RGB array is coated with a polymer that is designed to (a) allow only light of blue bandwidths to be transmitted through the polymer—acting like a filter—and sensed by that pixel, and (b) attenuate all other wavelengths, that is, prevent them from being sensed by that pixel. Typical CFA filters attenuate unwanted wavelengths by only 1/10th of the value of the maximum transmittance of the filter. This lack of rejection of light outside of the wavelengths of interest degrades color fidelity to an unacceptable level for accurate color measurement.
Due to signal detection problems caused by pixel spacing and poor color fidelity, CFA-type sensors are not accurate enough for satisfactory determination of tooth shade.
Currently, most intraoral cameras include a sheath to cover the illuminating portion and/or image sensor. Conventional sheaths are disposable, so that they may be replaced if they accidentally or intentionally come in contact with a patient's mouth. By replacing a sheath between measurements on different patients, a dentist may prevent spread of contaminants, such as infectious agents, from a first patient to subsequent patient. Although these protective sheaths prevent spread of contaminants, their functionality is limited exclusively to this sanitary purpose.
Conventional intraoral cameras also include a handheld probe that a dentist inserts into a patient's mouth and collects color images with. Via a cable, the probe transmits collected color measurement to a computer that subsequently processes the measurements to create images and displays those images on a monitor for the dentist to view. The drawback in collecting images of a tooth with these conventional probes is that the dentist must look back and forth from the probe to the monitor to insure the probe is positioned over the tooth to obtain the desired image on the monitor. This, of course can cause unneeded frustration in aligning the probe to collect measurements of the tooth.
In many instances, intraoral cameras or parts thereof intentionally or accidentally come into contact with a patient's intraoral cavity thereby transmitting contaminants including infectious agents, saliva and/or food debris to the device. In addition to using sanitary sheaths as discussed above, operators of prior art intraoral cameras frequently clean or sterilize the cameras. This is often a tedious task, as the cameras include a plurality of buttons that are difficult to clean around and/or fiber optic bundles that are nearly impossible to sterilize without damaging the optical characteristics of the fibers because sterilization agents enter the fiber optics and degrade illumination or sensing capabilities. Accordingly, prior art camera users must exercise time-consuming care in operating and cleaning these cameras.
Typically, a dentist makes a shade determination visually using shade tabs. A prescription describing the restoration location and shade is sent to the dental laboratory. There, a technician attempts to duplicate the tooth shade to make prosthesis from available ceramic or synthetic materials. Once the prosthesis is manufactured, it is sent back to the dentist for installation in the patient.
Only after placing the prosthesis in proximity to the patient's damaged tooth and/or surrounding teeth can the dentist determine if the prosthesis is an acceptable duplicate of the damaged tooth. Of course, if the prosthesis does not match properly, the dentist must have a second prosthesis made by a lab incorporating his suggested modifications. A second shade determination of the tooth may even be required. The second prosthesis must also be compared by the dentist to the damaged tooth to insure a proper match. This process is very costly if multiple prosthetic replacements must be produced to create a satisfactory match. Moreover, this process consumes the time of patients who may come in for repeated visits before a matching prosthesis is created.